The cross-searchability that Art is suggesting would depend on
consistency in naming fields. Do we need a standard for field names for
DC, MODS, MARC, etc.? Or should we abstract the fields; we could give
them numbers instead of names, and call them "use attributes"... yeah...


-----Original Message-----
From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of
Art Rhyno
Sent: Wednesday, May 31, 2006 12:06 PM
To: [log in to unmask]
Subject: Re: [CODE4LIB] fun with kinosearch

>The developer of KinoSearch and the
>developer of Ferret (the Lucene port for Ruby) are teaming with one of
>the developers of Lucene to make Lucy (which is a 'shared core' for
>Lucene) which would then make it easier to write interoperable code
>between these languages (at least, I presume that's the goal).

Lucene seems to be emerging as a common index format for many languages,
platforms, and projects. The availability of a common index format would
have some really nifty possibilities for libraries. For example, if
there were a lucene index for a citation database and the library's
holdings were also indexed by lucene, it might be possible to leverage
the same format to limit searches to materials that are immediately
available, rather than the "click and hope" model offered by resolvers
now. Or even arbitrary mixing and matching of databases at the index
level without having to replicate the content locally. My only quibble
with lucene is indexing throughput for very large databases but I am
told it can accommodate massive parallelization, in which case there
might be a new purpose for all of those library workstations that sit
empty in the wee hours...